Describe a Cross Sectional Study:
- A sample of, or total reference population is examined at a given point in time
- Records information on disease/outcomes + exposures at a single point in time
= snapshot of cohort study - Comparison of exposure in those with/without disease is equivalent to a case-control study
- Cross sectional sample is like sampling those with /without disease at one point in time
More about a Cross Sectional study
Can be population based or community based survey
Subjects are interviewed at a point in time without follow up
Can combine a cross sectional study with follow up to create a cohort study
Can conduct repeated cross sectional studies to measure changes against time
Point vs Period Prevalence
Point Prevalence : do you currently have a backache?
Period Prevalence : Have you had a backache in the last 6 months?
Non-equal probability sampling
Oversampling of particular populations in order to ensure adequate representation of those individual populations for analysis
Leads to a sample that doesn’t represent the general population
Survey Weight + Calculation
Survey weights are used to compensate of over-/under- sampling of subject groups
Ex: If we double the size of our sample from minorities, each minority gets a weight of 0.5
- Makes the statistics representative of the population
- Only 1 weight per subject. Weights for different factors must be combined into one weight
Calculation:
Survey weight = 1 / probability of sampling, or “sampling fraction,” or oversampling amount for a given group
Ex: oversample 5x Asians than Whites –> weight of Asians = 1/5, weight of Whites = 1
Bias in Cross Sectional Studies: Length-biased Sampling, Prevalence- incidence Bias
- Study population often accrued through convenience sampling (non-probability) = not representative of general population
- Systematic increase/decrease of cases
- length- biased sampling - cases are overrepresented if illness has a long duration, underrepresented if short duration - Systematic increase or decrease of exposure?
- prevalence-incidence bias: exposure doesn’t alter disease risk, but alters disease duration
Strengths of Cross Sectional Studies
- Relatively feasible, non-time consuming - no follow/up
- Can study several diseases/exposures, thus useful for screening new hypotheses
- We can describe disease frequency and health needs of a large population = useful for health planning
- Serial surveys can be done so that we can monitor a trend of disease
Limitations of Cross Sectional Studies
- Rarely know when disease occurred among cases = temporal ambiguity, did E precede O?
- Certain exposure info is vulnerable to measurement error, especially if collected retrospectively = bias in effect estimation
- Prevalence cases occurred before study started = disease status can influence selection of subjects
= bias in effect estimation - We don’t know duration of prevalent cases = can’t distinguish risk factors from prognostic factors
- Inefficient for rare/ highly fatal / short duration diseases
